Introduction

Infertility has emerged as a global health concern, impacting around 8-12% of couples during their reproductive years, approximately 50% of infertility cases are attributed to male factors (1). Among these, asthenoteratozoospermia, characterized by poor sperm motility and notable morphological abnormalities, stands out as a prevalent contributor to male infertility (2, 3). Recent investigations have unveiled a multitude of gene mutations associated with asthenoteratozoospermia and male infertility, including Dynein Axonemal Heavy Chain (DNAH) inner dynein arm (IDA) genes like DNAH1, DNAH2, DNAH6, DNAH7, and DNAH10, as well as outer dynein arm (ODA) genes such as DNAH8 and DNAH17, all implicated in primary male infertility with asthenoteratozoospermia (4–11). Additionally, genes encoding cilia and flagella-associated proteins (CFAP), including CFAP43, CFAP44, CFAP57, and CFAP61, have been identified as linked to asthenoteratozoospermia (12–15). Despite these advances, the intricate and heterogeneous features of asthenoteratozoospermia leave a significant number of cases unexplained. It is anticipated that additional genetic factors play a role in sperm anomalies and dysfunction.

The axoneme, a fundamental structure of cilia and flagella, exhibits a well-organized pattern with nine peripheral doublet microtubules (DMTs) and a central pair of microtubules (CP), constituting the classical “9+2” microtubular arrangement (16). In motile cilia or flagella, the radial spoke (RS) head contacts the CP accessory structures, transmitting mechanical and chemical signals from the central microtubules to the axonemal dynein arms, thus coordinating the flagellar or ciliary movement (17). Within mammalian species, the DNAH family plays a crucial role in the axonemal structure, encompassing eight IDA members (DNAH1-3, DNAH6, DNAH7, DNAH10, DNAH12, DNAH14) and five ODA members (DNAH5, DNAH8, DNAH9, DNAH11, DNAH17). Despite the collaborative effort of IDAs and ODAs in generating the force necessary for flagellar and ciliary motion, they serve distinct functions. Chlamydomonas ODAs, composed of α, β, and γ chains, primarily regulate beat frequency, while IDAs, with seven major subspecies (a-g), contribute to bending formation and beating form (18). Mutations in genes encoding ODA or IDA components have been linked to asthenoteratozoospermia, with or without symptoms of primary ciliary dyskinesia (PCD), a genetic disorder characterized by chronic airway infections or situs inversus (19). Notably, deficiency in certain IDAs and ODAs coding genes, such as DNAH7 or DNAH11, may lead to PCD (20–22). Patients carrying mutations in DNAH1 and DNAH2 typically lack PCD symptoms (4, 23), although a study suggested the potential contribution of the DNAH1 p.Lys1154Gln mutation to PCD (24). The extensive genetic heterogeneity and clinical variabilities underscore the intricacy and uniqueness of dynein arms in flagella and cilia. However, the susceptibility of uncharacterized dynein factors that related to asthenoteratozoospermia and/or PCD remains unclear, given the absence of human genetic evidence or corresponding animal models.

Dynein axonemal heavy chain 12, DNAH12 (MIM: 603340, GenBank: NM_001366028.2), also known as DHC3, is a unique dynein arm component (25). Notably, it is the shortest member and lacks a typical microtubule-binding domain (MTBD) in the DNAH family. Mutations of DNAH12 have been predicted to be associated with male patients’ infertility, however, limited pathogenic evidence was provided, and the underlying pathogenesis remained unclear (26, 27). Besides, although DNAH12 has been identified as a possible axonemal dynein and its mRNA is predominantly detected in two tissues containing axonemal structures: trachea and testis (25, 28), its physiological functions remain largely elusive.

In this study, we enrolled 796 infertile patients from unrelated Pakistani and Chinese families diagnosed with asthenozoospermia. Utilizing whole-exome sequencing (WES) followed by Sanger sequencing, we identified six novel bi-allelic mutations in DNAH12 in infertile males from six unrelated families. All these mutations resulted in the loss of DNAH12 protein in the sperm of patients. Two mouse models, Dnah12^−/−and Dnah12^mut/mut, were generated and exhibited asthenoteratozoospermia and male infertility, mimicking our patients’ phenotypes. Deficiency in DNAH12 led to axoneme defects characterized by impaired IDAs and CP loss in sperm flagella, while no such defects were observed in cilia. DNAH12 interacts with other IDA components like DNALI1 and DNAH1 in mouse testes, and disruption of DNAH12 impaired the recruitment of dynein components into sperm flagella. Our findings unveil, for the first time, that DNAH12 plays a pivotal role in facilitating the recruitment of IDA components specifically in flagella, but not in cilia. Furthermore, we demonstrate that bi-allelic mutations in DNAH12 are identified as causative factors for asthenoteratozoospermia and male infertility in both humans and mice. These results enhance our understanding of the development of axonemal structures in flagella versus cilia and provide valuable clues for genetic counseling and future individualized treatment for infertile males carrying DNAH12 mutations in clinical practice.

Materials and methods

Human subjects

Informed consent and peripheral blood were obtained from the participants and the patients’ information was registered in the Human Reproductive Disease Resource Bank database. Sperm analyses were performed according to the WHO guidelines (29). During their medical consultation, all patients answered a questionnaire regarding clinical presentations. This study was approved by the institutional ethics committee of University of Science and Technology of China (USTC) with the approval number 2019-KY-168.

Whole exome sequencing (WES) and bioinformatics analyses

Whole-exome capture and sequencing were performed by capturing exons from genomic DNAs of peripheral blood cells, using the SureSelect XT 50 Mb Exon Capture Kit (Agilent, California, USA), and then high-throughput sequencing was conducted on the HiSeq 2000 platform (Illumina, San Diego, CA, USA) following instructions. Variants were called using the Genome Analysis Toolkit HaplotypeCaller and annotated through ANNOVAR (30). Variants within exons or exon-intron boundaries were retained and filtered through the criteria we previously described (31). Bcftools was applied to detect runs of homozygosity (32). Runs of homozygosity over 1.5 Mb were used to calculate the inbreeding coefficients. The genotypes were validated using Sanger sequencing validation. Additional primers are listed in Supplemental Table 1. For phylogenetic analysis, different DNAH12 protein sequences were extracted from the UniProt database, and the sequence alignment and phylogenetic tree construction were generated using MEGA11 (33). Newick format tree files were extracted and processed in the Evolview v2 tool (34). For gene family and conserved domain analyses, the batch sequence search tool in Pfam was used to analyze features of proteins and further visualized with the assistance of Evolview v2 tool and TBtools (35, 36). For protein structure analysis, the structure models were predicted and constructed with the Phyre2 tool (37), interactions were analyzed by the HDOCK (http://hdock.phys.hust.edu.cn/) and Alphafold3 (https://golgi.sandbox.google.com/). The predicted protein structures and possible interactions were further explored and visualized by UCSF Chimera (Version: 1.15, California, USA) according to the manual instruction (38).

Scanning electron microscopy (SEM) and Transmission electron microscopy (TEM)

Fresh semen samples were obtained and centrifuged at 900 g for 3 min. Spermatozoa were washed with phosphate buffer saline (PBS) three times. Then the sediment was fixed in 0.1 M phosphate buffer (pH 7.4) containing 4% paraformaldehyde, 8% glutaraldehyde, and 0.2% picric acid at 4°C for at least 8 hours. Scanning electron microscopy (SEM) and Transmission electron microscopy (TEM) were then conducted as we previously described (39).

Generation and verification of the polyclonal anti-DNAH12 antibody

The characteristics of the DNAH12 were analyzed and the 1-200 amino acids sequence of mouse DNAH12 protein (UniProt accession no. Q69Z23) was regarded as the ideal antigen for immunization. Subsequently, the DNAH12 polyclonal antibody was generated in rabbits and rats using the mentioned peptide by Dia-An Biotech, Inc., in Wuhan, China. The specificity of antibodies was further validated using Dnah12^+/+ and Dnah12^−/− mice in this study.

Immunoblotting

Protein lysate was prepared as we previously described (40, 41). Tissues or cells were lysed using lysis buffer (50 mM tris (pH 7.5), 150 mM NaCl, 5 mM EDTA, and 0.1% NP-40). For human sperm cells, proteins were extracted using the TRIzol reagent. The protein sediment was washed with 95% ethanol containing 0.3M guanidine hydrochloride, dissolved in 1% SDS. Then the protein lysate was boiled for 10 minutes, and diluted in SDS-sample buffer (120 mM Tris/HCl, 10% SDS, 20% glycerol, 20% 2-mercaptoethanol, and bromophenol blue (pH 6.8). Whereafter the proteins were separated by SDS-PAGE and transferred onto nitrocellulose blotting membranes with a pore size of 0.45 μm (GE Healthcare, 10600002). The membranes were blocked with TBS-Tween buffer (50 mM Tris, 150 mM NaCl, and 0.5% Tween-20, pH 7.4) containing 5% skim milk for 30 minutes to reduce unspecific binding and then incubated with primary antibodies diluted in TBST buffer containing 5% skim milk at 4°C overnight. Incubation with secondary antibodies for 1.5 hours was followed by chemiluminescence development system (GE Healthcare, Pittsburgh, USA). Antibodies details are listed in Supplemental Table 2-3.

Immunofluorescence (IF) staining

Human semen smear slides were prepared following the guidelines of WHO (29). Mouse sperm smear slide preparation and immunofluorescence staining were conducted as we described (42). For tissue IF staining, tissues were collected and fixed in 4% paraformaldehyde (PFA), the 5μm tissue sections were deparaffinized in xylene, rehydrated in a graded series of ethanol (100%, 95%, 90%, 80%, 70%, and 50% ethanol) and sterile water, and permeabilized with 0.2% Triton X-100 in PBS for 30min. Antigen retrieval was then conducted by boiling slides in a microwave for 30 minutes in 10 mM citrate buffer (pH 6.0). The other steps are consistent with sperm IF staining. Images of spermatozoa were captured using a Nikon ECLIPSE 80i microscope equipped with a charge-coupled device (Hamamatsu Photonics, Hamamatsu, Japan). Information of antibodies is available in Supplemental Table 2-3.

Animal models and fertility test

Two mouse models (Dnah12^−/− and Dnah12^mut/mut) were generated using CRISPR/Cas9 Technology (43). Briefly, to generate the Dnah12^−/− mouse model, guide RNAs (gRNAs) targeting exon 5 of Dnah12, a key exon shared by all its transcripts, were transcribed in vitro (Thermo Fisher, CA, USA). Then, the gRNAs and the Cas9 protein were electroporated into C57BL/6 mouse zygotes. The zygotes were introduced into ICR female mice, and the genotypes of the pups were checked using Polymerase Chain Reaction (PCR) with designed primers. Similar conductions were performed targeting exon 18 of Dnah12 to generate the Dnah12^mut/mut mice. The genotyping primers are listed in Supplemental Table 1. For male fertility test, each 8-10 weeks male mouse was caged with wild-type female mice; for female fertility test, each 8-10 weeks female mouse was caged with one 8-10 weeks Dnah12^+/- male mouse, the number of offspring was recorded.

Reverse-transcription PCR and real-time quantitative PCR

For reverse-transcription PCR (RT-PCR), total RNA was extracted from mouse tissues using TRIzol reagent (TAKARA, Japan). Whereafter 1μg of RNA was reverse-transcribed using the PrimeScript RT Reagent Kit (TAKARA, Japan). The cDNA was diluted to a comparable concentration and then used for RT–PCR. All amplified DNA fragments were visualized by 2% agarose gel electrophoresis containing GelRed (APExBIO, Houston, TX, USA). For real-time quantitative PCR (qPCR), the experiments were performed by using Hieff UNICON® Universal Blue qPCR SYBR Green Master Mix reagent (Yeasen Biotechnology, Shanghai, China) on CFX96® Real-Time System (Bio-Rad, CA, USA) according to the manufacturer’s guidelines. Details of the primers were listed in Supplemental Table 1.

Histological analysis

As previously described (31), after sacrificing mice, the testes, epididymides, tracheas, and oviducts were harvested and fixed in Bouin’s fixative solution (Sigma–Aldrich, Missouri, USA) for hematoxylin and eosin (H&E) staining or 4% PFA fixative solution for immunofluorescence or TUNEL assay. The pictures were captured using a microscope equipped with a charge-coupled device camera (Nikon, Tokyo, Japan).

Sperm counts, motility and morphology analysis

As previously described (44), H&E staining or Papanicolaou staining kit (Solabio, Beijing, China) was used for the evaluation of sperm morphology of patients. For mouse sperm parameters analysis, 8-10-week-old mice were sacrificed by cervical dislocation, the sperm number per epididymis was then assessed by cutting the epididymis into pieces in PBS and incubating them for 30 minutes at 37°C to release the sperm, then counting with a hemocytometer. Sperm motility were assessed by computer-aided sperm analysis (CASA) as we described (14). Sperm morphology was assessed on sperms collected from epididymal cauda after H&E staining.

Testicular cell smear slides and seminiferous tubule squash slides preparation

To prepare testicular cell smears, mice were sacrificed and testes were then detached with one time of PBS washing. Then testes were peeled off the testicular capsule, torn into small pieces, and then cut into tiny pieces in PBS. Finally, the cell suspension was filtered through a 200-mesh cell strainer, the filtered cell suspensions were added 10% FBS before spread onto glass slides. For testicular seminiferous tubule squash slide preparation, the testes were dissected and carefully isolated into single testicular seminiferous tubules in PBS dishes. After 10 mins of prefixation in fix/lysis solution (2% PFA, 50% PBS), we then applied 100μL fix/lysis solution onto poly-L-lysine coated glass slides and carefully teased apart approximately 20 seminiferous tubules with the assistance of 2 sterile forceps on the slides. After that, coverslips were applied using slide forceps and pressed with the heel of the palm for 10 seconds, then the glass slides were frozen in a small container with liquid N₂ immediately till the liquid ceases to bubble. The slides were saved in the −80°C fridge keeping the coverslip on the slides or removing coverslips for immunofluorescence staining.

TUNEL assay

As previously described (44), In Situ Cell Death Detection Kit, Fluorescein (Sigma–Aldrich, Missouri, USA) was applied according to the manufacturer’s protocol. Images were captured using a Nikon Eclipse 80i microscope equipped with Hamamatsu C4742-80 digital camera and analyzed using the NIS-Element Microscope imaging software (Nikon, Tokyo, Japan).

Cell culture and transfection

As previously described (40), HEK293T (ATCC, VA, USA) cells were cultured to 70%-80% density in 12-well plates, followed by transfected with plasmids expressing mCherry-tagged mouse DNAH12 and GFP-tagged mouse DNALI1 using Lipofectamine 3000 (Thermo Fisher, CA, USA) according to the manufacturer’s instructions. Cells were harvested for in-vitro IP and immunoblotting the next day after transfection. The sequences of the primers for plasmid construction are listed in Supplemental Table 4.

Co-immunoprecipitation (Co-IP)

Tissue extracts were prepared using lysis buffer supplemented with PMSF, Protease Inhibitor Cocktail (Solarbio, Beijing, China) on ice using a glass homogenizer. Lysate was then centrifuged at 12,000 g at 4°C for 10 min, and the supernatant was divided into equal aliquots. Each aliquot was incubated with 1.5 μg of target antibody or rabbit IgG nonspecific antibody. After incubation at 4°C overnight with rotation, beads were washed five times with lysis buffer. Finally, the beads were resuspended in SDS-sample buffer and boiled at 100°C for 10 min. The samples were either analyzed by immunoblotting immediately or stored at −80°C. For Co-IP in cultured cells, HEK293T cells were plated in six-well plates, followed by transfection using Lipofectamine 3000 reagents. MG132 was added into the cultured cells 6 hours before cell harvesting. Cells were washed with PBS for 10 min, cells were then collected and lysed in the same lysis buffer used for Co-IP from testicular samples. The cell lysate was then incubated with prewashed Magnetic Beads (Bimake, Texas, USA) and primary antibodies at 4°C overnight, followed by washing and boiling as described above for the Co-IP protocols for testicular samples.

Mass spectrometry proteomics

To identify potential partners that interact with DNAH12, we performed in vivo IP of mouse testes lysate. Tissue extracts were prepared using lysis buffer supplemented with PMSF, Protease Inhibitor Cocktail (Solarbio, Beijing, China) with the same operation of the Co-IP protocol. Beads were washed 3 times with lysis buffer and 3 times with 50 mM NH₄HCO₃, the pull-down proteins were digested with 0.2μg Trypsin overnight and stopped by 5% Formic acid, (Thermo Fisher, CA, USA). The peptide mixture was then desalted, dried with a SpeedVac, and resuspended in 0.1% Formic acid, followed by the MS analysis using timsTOF Pro 2 mass spectrometer (Bruker, Massachusetts, USA). The proteomics data are available on the ProteomeXchange Consortium via the iProX partner repository (45), with the dataset identifier PXD051681.

Intracytoplasmic sperm injection (ICSI)

3.5-4-week-old female mice were given an intraperitoneal injection of 5 IU Pregnant Mare Serum Gonadotropin (PMSG, Ningbo Sansheng, China). After 48 hours, a second intraperitoneal injection of 5 IU human Chorionic Gonadotropin (hCG, Ningbo Sansheng, China) was administered. The mice were euthanized by cervical dislocation after 15 hours post-hCG injection. The oviducts were excised and the ampullary regions were opened to release the cumulus-oocyte complexes (COCs). The cumulus cells surrounding the oocytes were digested using hyaluronidase, and the morphologically round MII stage oocytes were selected and washed three times in M2 medium (Sigma–Aldrich, Missouri, USA) before being set aside for use. Male mice are euthanized by cervical dislocation, the epididymides are excised and gently punctured with a syringe, allowing the spermatozoa to flow out. The spermatozoa were then placed in the human tubal fluid (HTF) medium (EasyCheck, Nanjing Aibei Biotechnology, China) and then frozen and thawed repeatedly to remove sperm tails, the single sperm head was microinjected into an MII oocyte and then cultured in potassium simplex optimized medium (KSOM, EasyCheck, Nanjing Aibei Biotechnology, China) at 37 °C under 5% CO_2.

Statistics analysis

Statistical analyses were performed using GraphPad Prism 8.0 (GraphPad Software, San Diego, CA, USA). Values in tables and graphs are expressed as mean ± SEM. Two groups of mice were compared with Student’s 2-tailed t-test for independent data and P > 0.05 was considered not significant (ns); *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 were considered significant.

Results

Identification of biallelic DNAH12 variants in men with asthenoteratozoospermia

To identify the genetic factors contributing to male infertility, we recruited patients with primary infertility and their family members from Pakistan and China. WES was performed on these individuals, followed by comprehensive analyses of the WES data (Supplemental Figure 1, A-D). Seven mutations in DNAH12 were identified as potential pathogenic factors for male infertility in patients P1-3 of family 1, P4-5 of family 2, P8 of family 3, and P9, P10, and P11, all of whom suffered from asthenoteratozoospermia (Fig. 1A). Initially, a homozygous frameshift mutation c.944_945 del (p. Phe315*) in family 1, and two homozygous missense mutations, c. A164G (p. Gln55Arg) in family 2 and c. A2560G (p. Ile854Val) in family 3, were identified in DNAH12 as potential causes of infertility. Subsequently, four other mutations in DNAH12 were discovered in three sporadic Chinese patients with asthenoteratozoospermia: a homozygous stop-gain mutation c.2680 C>T (p. Gln2077*) in subject P9, a homozygous missense mutation c.2813T>C (p. Val938Ala) in P10, and compound heterozygous mutations c.2680C>T (p. Arg894Cys) and c.5964-3A>G in P11 (Fig. 1A and B).

Figure 1.

All seven variants were rare in public human genome databases, including the 1000 Genome Project and gnomAD (v3). Additionally, these variants were predicted to be deleterious by in-silico software tools such as SIFT, Polyphen2, fathmm-MKL, and CADD2 (Table 1). Sanger Sequencing further validated these variants and revealed the recessive pathogenic pattern of these mutations (Fig. 1A). Collectively, these results suggested that the identified bi-allelic DNAH12 variants could potentially be the cause of asthenoteratozoospermia in our patients.

Table 1.

DNAH12 protein is conserved in mammals and highly expressed in the testes

DNAH12 encodes a functionally uncharacterized protein and is predicted to be abundantly expressed in testes of humans and mice based on our bioinformatics tool FertilityOnline (46). Experimentally, our RT-PCR, qPCR, and immunoblotting detection of DNAH12 expression in different tissues indicated that both DNAH12 mRNA and protein showed high expression in testes and relatively lower expression in lungs and tracheas of male mice and oviducts of female mice (Supplemental Figure 2, A-C). It’s worth noting that in epididymides, DNAH12 protein, but not mRNA, was detected, indicating the DNAH12 protein detected in epididymides is from testes, most probably from sperm. Furthermore, DNAH12 protein was initially detected in testes around 21 days postpartum (dpp) and thereafter maintained a high level (Supplemental Figure 2D). The expression pattern of DNAH12 in testes suggests that it is possibly involved in spermiogenesis, and its abnormal expression could lead to asthenozoospermia or asthenoteratozoospermia, further indicating that the identified DNAH12 mutations are candidate pathogenic mutations for patients.

Further phylogenetic analysis revealed the conservative nature of DNAH12 from Chlamydomonas to humans (Supplemental Figure 3). All the amino acids affected by mutations were conserved in mammals and were located in functional regions, including the ATPase domains or the Stem domain of DNAH12 (Fig.1C), indicating that these mutations are highly likely to hinder the structure, function, or stability of the DNAH12 protein. Altogether, our results suggested that the identified mutations in DNAH12 are probable pathogenic factors contributing to asthenoteratozoospermia in our patients.

Impaired sperm motility and morphology in patients with bi-allelic DNAH12 variants

To assess the pathological effects of the identified bi-allelic DNAH12 variants, semen analyses were conducted following WHO guidelines (29). All affected individuals exhibited significantly reduced sperm motility and progressive motility (Table 2). Detailed sperm morphology analyses were performed on patients P3 from family 1, P5 from family 2, P8 from family 3, and three Chinese sporadic patients (P9, P10, and P11). Spermatozoa from fertile controls displayed normal heads and tails, while spermatozoa from the patients exhibited a range of abnormal morphologies, including short, coiled, absent, bent, or irregular caliber flagella, accompanied by aberrant heads (Fig. 2A, Supplemental Figure 4, A and B). Scanning electron microscopy (SEM) analyses were further carried out on a fertile control and P10, confirming the presence of abnormal morphologies of the sperm head and flagellum in the patients (Fig. 2B). Thus, patients with the bi-allelic DNAH12 variants demonstrated significantly decreased sperm motility and increased abnormal sperm morphology.

Figure 2.

Table 2.

The bi-allelic DNAH12 variants result in absence of DNAH12 protein in patients’ sperm

To investigate the effects of the identified bi-allelic DNAH12 variants on the DNAH12 protein, we obtained sperm samples from three affected Chinese patients, and the expression of DNAH12 was detected using a custom-made DNAH12 rabbit antibody that recognized the N-terminal 1-200 amino acids of the DNAH12 protein. This validated antibody (Supplemental Figure 6B) could theoretically detect the truncated protein if it were present in P9 with a homozygous stop-gain mutation M4 (p. Gln2077*). The predicted DNAH12 protein size was detected in the sperm lysate from a fertile control but not in P9 with the homozygous stop-gain mutation M4, P10 with a homozygous missense mutation M5, and P11 with compound heterozygous mutations M6 and M7, indicating these mutations possibly lead to the disruption of the DNAH12 protein (Fig. 2C).

To further explore the effects of the identified DNAH12 variants on the DNAH12 protein, we performed immunofluorescence assays on sperm from patients. The results showed that DNAH12 signals were localized along the flagella of the fertile control, however, no DNAH12 signals were observed in sperm flagella of patients P3 (family 1), P5 (family 2), P8 (family 3), P9, P10, and P11 (Fig. 2D). Collectively, these findings indicated that all the detected DNAH12 variants cause the absence of DNAH12 protein in the sperm of our patients, reinforcing that these DNAH12 mutations are pathogenic mutations of asthenoteratozoospermia in the patients.

Patients with bi-allelic DNAH12 variants exhibit flagellar axoneme defects

Since patients with the bi-allelic DNAH12 variants exhibited compromised sperm motility and abnormal sperm morphology, we then investigated whether sperm flagellar ultrastructure was also affected in these patients. TEM examination was performed on sperm flagella from both patients and normal fertile controls. The flagella of fertile controls displayed the typical “9+2” microtubule structure with well-organized DMTs and CP. In contrast, various defects were found in our patients, including loss of CP (54.5%), disorganization of microtubules (26%), and absence of DMTs (16.5%) in the examined cross-sections (Fig. 3A, 3C and Supplemental Figure 5A). To examine the IDAs impairment in axonemes, the sections with intact ODAs were analyzed. Compared with axonemes in controls that almost all the sections presented intact IDAs, approximately 70% of sections examined showed impaired IDA structures in the analyzed three patients (Fig. 3B, 3D and Supplemental Figure 5A).

Figure 3.

We then performed immunofluorescence assays on sperm smears to assess the expression and localization of specific proteins of the axonemes. These included SPAG6 (a CP component), DNAH1 and DNALI1 (IDA components), DNAH17 (an ODA component), and ODF2 (an outer dense fiber, ODF component). The sperm flagella of patients showed remarkably reduced SPAG6 signals when compared with those from controls (Fig. 3E). Similarly, DNAH1 and DNALI1 signals were both obviously decreased in the patients’ sperm flagella, indicating impaired IDAs in the patients (Fig. 3F and G). Besides, DNAH17 and ODF2 signals in patients’ flagella were comparable to those in a normal control (Fig. 3H and Supplemental Figure 5B). In summary, these findings indicated that the bi-allelic DNAH12 variants cause pronounced flagellar axoneme defects, including impaired IDAs and loss of CP, but do not affect ODAs and ODFs.

Deletion of Dnah12 impairs spermiogenesis and causes male infertility in mice

Considering the bi-allelic DNAH12 variants resulting in absence of DNAH12 protein in patients’ sperm, to functionally confirm the deletion of DNAH12, we generated Dnah12 knockout mice using CRISPR-Cas9 approach (Supplemental Figure 6A). Fortunately, we obtained a mouse model with a 4-bp deletion (c.386_389del) in exon 5 of Dnah12 (Supplemental Figure 6A). In contrast to the observations in control mice, immunoblotting of testis lysate using two antibodies targeting the N-terminal of DNAH12 failed to detect DNAH12 proteins, and immunofluorescence staining of sperm flagella did not show DNAH12 signals in Dnah12^−/− mice (Supplemental Figure 6, B and C). Dnah12^−/−males exhibited normal development but were infertile (Fig. 4A), while Dnah12^−/− female mice showed normal development and fertility, and no obvious defective oviductal and tracheal cilia (Supplemental Figure 7, A-C). The body weight of Dnah12^−/− and Dnah12^+/+males was comparable, while smaller testes were observed in Dnah12^−/−males (Fig. 4B). Besides, the average number of sperm per epididymis of Dnah12^−/−mice dramatically decreased to approximately 2% of that of Dnah12^+/+ mice (Fig. 4C). The epididymal caudal spermatozoa of Dnah12^−/− male mice were 100% immotile, compared to those of Dnah12^+/+ mice (Fig. 4D, Supplemental Videos 1 and 2). H&E staining of testis and epididymis sections also showed a striking reduction in the number of sperm in Dnah12^−/− mice (Fig. 4E). Importantly, morphological analysis of sperm from the caudal epididymis of Dnah12^−/−mice showed similar types and frequencies of flagellum and head abnormalities to those observed in our patients (Fig. 4F and G).

Figure 4.

To understand how DNAH12 depletion causes abnormal sperm morphology, we conducted immunofluorescence assays on testicular cells from Dnah12^+/+and Dnah12^−/− mice, in Dnah12^−/− mice, while DNAH12 signals are missing in round, elongating, and elongated spermatids, the abnormal manchette structure and flagellum were observed (Fig. 4H). We then performed immunofluorescence assays on testicular sections using α-Tubulin and lectin peanut agglutinin (PNA, a marker of the acrosome) to examine the process of spermiogenesis. The manchette structure in Dnah12^−/− was roughly normal until steps 9-10 (Stage IX-X) but was aberrantly stretched from steps 11-12 (Stage XI-XII) when compared with those of Dnah12^+/+ mice (Supplemental Figure 8). To meticulously characterize these defects, we stained α-Tubulin and PNA on squashed seminiferous tubules and found obvious manchette defects and abnormal sperm heads in Dnah12^−/− mice (Fig. 4I). TEM analysis showed severely abnormal head morphology in elongated spermatids of Dnah12^−/− males when compared with those of Dnah12^+/+ males (Fig. 4J). Additionally, TUNEL assays of testis sections showed dramatically increased proportions of seminiferous tubules containing TUNEL-positive cells in Dnah12^−/− mice, compared with those in Dnah12^+/+ mice, suggesting the aberrant spermatids may undergo apoptosis (Figure 4K-M). Altogether, these results indicated that deletion of Dnah12 impairs spermiogenesis, results in decreased sperm counts, and increased sperm morphological abnormalities in mice.

Deletion of Dnah12 causes axoneme defects in sperm flagella but not cilia in mice

TEM analysis was conducted on sperm from cauda epididymis of Dnah12^+/+, Dnah12^+/- and Dnah12^−/− mice. Flagellar axonemes of Dnah12^+/+and Dnah12^+/- mice displayed a typical “9+2” microtubular arrangement with clearly visible CP, IDAs, and ODAs. In contrast, sperm from Dnah12^−/− mice exhibited severe abnormalities in all the analyzed sections, characterized by CP loss and disorganized axoneme structure (Fig. 5A). Immunoblotting was performed to assess the expression levels of DNAH1, DNALI1, and SPAG6 in sperm lysates, Dnah12^−/−sperm showed a significant reduction in the abundance of DNAH1 and DNALI1, while the amount of DNAI2 (a component of ODAs) remained comparable to Dnah12^+/+and Dnah12^+/- sperm (Fig. 5B). Additionally, immunofluorescence assays revealed the near absence of signals for SPAG6, SPEF2 (a component of CP complex), DNAH1, and DNALI1 in Dnah12^−/− sperm flagella, whereas DNAH17 and DNAI2 (components of ODAs) remained unaffected (Fig. 5C-H). These results indicate that DNAH12 deficiency specifically impairs CP and IDAs, but not ODAs, in sperm flagella. Furthermore, immunoblotting of SPAG6 and DNALI1 in testis lysate exhibited a sharp decrease in Dnah12^−/− mice (Supplemental Figure 9A). Additionally, TEM examination of testicular sperm axonemes showed defects like loss of CP and disarranged axonemes in Dnah12^−/− mice (Supplemental Figure 9B). These results indicate that impaired IDAs and loss of CP possibly occur during testicular sperm flagellar development.

Figure 5.

Another Dnah12 mutant mouse model (Dnah12^mut/mut) targeting exon18 was also generated (Supplemental Figure 10A), which may produce a truncated DNAH12 protein (p.Cys793Glu, fs*10). The mutation effect was checked but no truncated DNAH12 protein was detected, indicating that DNAH12 expression was disrupted in Dnah12^mut/mut mice (Supplemental Figure 10B). Moreover, Dnah12^mut/mut mice developed normally, while decreased testicular size was observed (Supplemental Figure 10, C and D). Consistent with Dnah12^−/− mice and patients carrying bi-allelic DNAH12 mutations, Dnah12^mut/mut presented obviously decreased sperm count, increased sperm morphological abnormalities, and sperm axoneme defects similar to Dnah12^−/− mice (Supplemental Figure 10, E-H). Therefore, these results reconfirmed that DNAH12 is essential for sperm flagellar development.

Given the detection of DNAH12 protein in the trachea and lung (Supplemental Figure 2, A-C), we also examined whether the deficiency of DNAH12 would cause PCD symptoms in males. Intriguingly, both Dnah12^mut/mut and Dnah12^−/− males developed normally and showed no tracheal defects, despite the expression of DNAH12 being detected in tracheal cilia cells (Supplemental Figure 11, A-E). These findings collectively suggested that DNAH12 deficiency is crucial for sperm axonemal integrity in humans and mice without manifesting PCD symptoms.

DNAH12 interacts with IDA components DNAH1 and DNALI1

Through phylogenetic and domain analyses of DNAH family members, we found that DNAH12 is a distinctive IDA lacking the microtubule-binding domain (MTBD), which is crucial for ATP-sensitive microtubule binding and motor direction control (Fig.6A, Supplemental Figure 12). Therefore, DNAH12 likely collaborates with other members to carry out its functions. To delve into the molecular mechanism of DNAH12, we conducted co-immunoprecipitation (Co-IP) followed by mass spectrometry analysis using a DNAH12 rabbit antibody in testicular lysate from adult Dnah12^+/+ mice (Fig. 6B). Gene ontology analysis showed that the potential interactors were mainly enriched in motor protein, actin binding, myosin and RNA binding proteins (Fig. 6C). The motor proteins were predominantly categorized into three groups: dynein, kinesin, and myosin protein members, interestingly, among these, dynein proteins such as DNAH1, DNALI1, DNAH1, DNAH3, and DNAH2 were annotated to be involved in IDA assembly, which is closely related to the defects found in DNAH12 mutants (Fig. 6D). This piqued our interest to explore these proteins in detail. Besides the predominant enrichment of DNAH12 (top1), potential IDA interactors such as DNALI1 (top2), DNAH1 (top3), DNAH3 (top4), and DNAH2 (top5) were co-enriched in the candidate interacting dynein protein list (Supplemental Table 5). Immunoblotting confirmed the co-enrichment of DNAH12, DNALI1, and DNAH1 in the immunoprecipitated complex by the DNAH12 rabbit antibody in Dnah12^+/+mice but not in Dnah12^−/− mice, indicating specific interactions between DNAH12 and DNALI1, DNAH1 (Fig. 6E). Reciprocal in vivo Co-IP using an anti-DNALI1 antibody further validated strong interactions among these three IDA components (Fig. 6F).

Figure 6.

Given the potential role of DNALI1 as an IDA assembly factor (47), we delved into the interaction details between DNAH12 and DNALI1. Employing Phyre2, we constructed the structural models of both proteins and further elucidated their interaction dynamics via DNAH12-DNALI1 docking using the HDOCK server. The results showed that the N-terminal Stem domain of DNAH12 possibly interacts with DNALI1 in the top three best-fit conformations (Fig. 6G). Considering DNAH12 is encoded by 3092 amino acids and their full-length protein interactions have been proved in vivo, we constructed the Stem domain (1-1212 amino acids) of DNAH12 and full-length DNALI1 plasmid, further in vitro Co-IP assays validated that the Stem domain of DNAH12 indeed mediates interactions with DNALI1 (Fig. 6H-J). Together, these findings proved the specific interactions among DNAH12, DNALI1, and DNAH1. Given the significantly decreased protein levels of DNALI1 and DNAH1 observed in the sperm of Dnah12^−/−mice (Fig.5B), we proposed that the interactions involving DNAH12 (possibly restricted to the Stem domain of DNAH12), DNALI1, and DNAH1 are crucial for flagella development. As dynein proteins typically share similar domains, our findings also suggested that the Stem domain of DNAH family members may facilitate the interaction with DNALI1.

Deletion of DNAH12 impairs the recruitment of IDA components during flagella development

To evaluate the effect of DNAH12 deletion on DNALI1 and DNAH1 during spermiogenesis, we examined the abundance of the DNAH12 interaction partners, DNALI1 and DNAH1, in Dnah12^+/+ and Dnah12^−/− mouse testis lysate. The protein levels of DNALI1 and DNAH1 were decreased in Dnah12^−/−testes, while the level of ODA member DNAI1 was comparable (Fig. 7A). We then stained DNALI1 or DNAH1 on testicular cell smear of Dnah12^+/+and Dnah12^−/− mice. Although the signals of DNALI1 or DNAH1 in round spermatids of Dnah12^−/− mice were comparable to those of Dnah12^+/+ mice (Fig. 7B and 7C), the recruitment of DNALI1 and DNAH1 to sperm flagella was severely impaired in elongated spermatids of Dnah12^−/− mice (Fig.7D and 7E). Considering that the deficiency of DNAH12 also caused scarcely visible immunofluorescence signals of DNALI1 and DNAH1 in sperm flagella of humans and mice (Fig. 3F, 3G, 5E, and 5F), we proposed that the interaction among DNAH12, DNALI1, and DNAH1 is conserved in mammals, and is required for flagellar development during spermiogenesis. Since the deficiency of DNAH12 does not cause PCD symptoms in humans and mice, we hypothesized that flagella and cilia may exhibit distinct features in the process of dynein assembly. We thus checked the abundance of DNALI1 and DNAH1 in the oviduct and trachea, the results showed that DNALI1 and DNAH1 were comparable in the Dnah12^+/+ and Dnah12^−/− groups (Fig. 7F and 7G). Furthermore, no discernible disparities were observed in the intensity and localization of DNALI1 and DNAH1 signals within the oviduct and trachea (Fig. 7H-K). In summary, these results underscore the specific involvement of DNAH12 in flagellar development, rather than in cilial development. Moreover, we supposed that the assembly of IDA components differs between flagella and cilia.

Figure 7.

DNAH12 is prerequisite for the recruitment of DNALI1 and DNAH1 to sperm flagella

In patients and mice with DNALI1 deficiency, flagellar DNAH1 but not DNAH12 signals were absent (47). To unravel the recruitment relationships among DNAH12, DNALI1, and DNAH1 during sperm flagellar formation, we conducted staining of DNALI1 and DNAH12 on sperm flagella obtained from the infertile asthenoteratozoospermic male P13 with a homozygous DNAH1 mutation (c.11275C>T) of a family (Fig. 7L, and Supplemental Figure13 A and B), wherein DNAH1 signals were absent along the flagella (Fig. 7M). In comparison to those in the control, the DNALI1 signals were absent, while no discernible differences in DNAH12 signals were observed in the flagella of the patient with the DNAH1 mutation (Fig. 7N and 7O). These findings suggested that the recruitment of DNAH12 to flagella is independent of DNAH1, and DNAH12 is the first protein to be recruited to flagella among the three proteins.

DNAH12 deficiency causes the decrease of RS head associated proteins

In addition to the enrichment of dynein proteins, we noticed that a clusters of RS head-related proteins such as RSPH1, RSPH9, or DNAJB13 were enriched in the testicular interactome of DNAH12. Considering that deficiency in RSPH1 (48), RSPH9 (49), or DNAJB13 (50) were associated with CP loss phenotype. We then explored their interaction with DNAH12 in the testis lysate, the results proved the interactions of these proteins with DNAH12 in the testis (Fig. 8A). Further analyses by AlphaFold3 showed that these proteins may interacts with DNAH12 through Stem domain (Fig. 8B). The expression of RSPH1, RSPH9, and DNAJB13 showed obvious decrease in testis lysate of Dnah12^−/− mice (Fig. 8C). Additionally, immunofluorescence assays revealed the loss of signals for RSPH1, a marker of RS head, in Dnah12^−/− sperm flagella (Fig. 8D). Together, these results indicate that deficiency of DNAH12 lead to the impairment of radial spoke head, which could contribute to CP loss in flagella of Dnah12^−/− mice.

Figure 8.

ICSI may overcome the infertility caused by DNAH12 deficiency

Since the DNAH12 mutated patients and the Dnah12^−/−mice have similar deficiency, to evaluate the prospects of whether the asthenoteratozoospermia patients caused by DNAH12 variants could be treated with assisted reproductive technology, spermatozoa from Dnah12^+/+and Dnah12^−/− mice were used for ICSI. Promisingly, the percentages of 2-cell and blastocyst-stage embryos in Dnah12^−/− group showed no difference compared to those of Dnah12^+/+ group (Fig. 8E-G), indicating that ICSI could also be an effective treatment for infertile men carrying DNAH12 variants.

Taken together, we suggest that DNAH12 interacts with DNALI1 and DNAH1, and this interaction is a prerequisite for the proper localization of DNALI1 and DNAH1 to sperm flagella, which is essential for the intact IDAs and CP structures in both humans and mice. Additionally, DNALI1 and DNAH1 exhibit mutual dependence for their localization to sperm flagella. Deficiency of DNAH12 leads to the failure of DNALI1 and DNAH1 recruitment to sperm flagella, resulting in abnormal sperm morphology characterized by compromised axoneme organization, causing male infertility (Fig. 8H). Moreover, our results indicate that the assembly of IDA components occurs intricately, possibly not synchronously in time and space.

Discussion

In this study, we identified six novel bi-allelic mutations in DNAH12 that cause the loss of DNAH12 proteins in asthenoteratozoospermia patients. Spermatozoa from these patients exhibited abnormal morphology and compromised axonemal defects, including a high percentage of impaired IDAs and loss of CP. Two mouse models (Dnah12^−/− and Dnah12^mut/mut) were generated, recapitulating the abnormal flagellum morphology and structure observed in patients, thereby confirming the pathogenic role of DNAH12 mutations in asthenoteratozoospermia. Through testicular Co-IP and mass spectrometry proteomics, we screened and validated that DNAH12 interacts with IDA components like DNALI1 and DNAH1 and is indispensable for recruiting DNALI1 and DNAH1 to sperm flagella, highlighting the crucial role of DNAH12 in sperm flagellar development. Notably, deficiency of DNAH12 does not manifest PCD symptoms in humans and mice, underscoring the unique role of DNAH12 in flagellar development. Collectively, our study not only unveils the functional significance of the previously unexplored gene DNAH12 but also elucidates the pathogenic link between DNAH12 mutations and asthenoteratozoospermia.

DNAH12 and other dynein arm proteins share the N-terminal Stem domain, which is crucial for structure and accessory subunit binding (51), and the motor domain which encompasses ∼6 AAA+ ATPase modules forming the ring-shaped oligomeric complexes responsible for force production (52). Five of the identified DNAH12 mutations, primarily in the Stem domain (critical for DNALI1 interaction), result in loss-of-function effects, indicating a weak tolerance property of the Stem domain. Similar weak tolerance features were observed in mutations of other dynein arm components like DNAH1 and DNAH2 (53, 54). The profound impact of these variants, particularly missense mutations, underscores delicate features of the Stem domain and DNAH family members. Notably, DNAH12 stands out as the shortest in sequence length and the only one lacking the essential MTBD domain within the DNAH family (55). MTBD, a 15-nanometer coiled-coil stalk structure, crucially supports ATP-sensitive microtubule binding properties and determines dynein motility direction (56). Our investigation revealed that DNAH12 interacts with other IDA components, specifically DNALI1 and DNAH1. Despite the mutual dependence between DNAH1 and DNALI1 for their flagellar localization, DNAH12 is prerequisite for the proper assembly of DNALI1 and DNAH1 onto sperm flagella in both humans and mice. Notably, Dnah12^−/−mice exhibit more severe phenotypes than Dnali1 and Dnah1 knockout mice (4, 47, 57), including a more severe reduced sperm count (∼ 2% of normal values), this observation further supports our perspective.

Deficiency in DNAH12 presented sperm ultrastructural defects like CP loss and impaired IDAs indicating essential roles of DNAH12 in the maintenance of axonemal integrity and arrangement. These results indicate essential roles of DNAH12 in the maintenance of axonemal integrity and arrangement. Noteworthy, deficiency of DNAH12 potential partner DNAH1, a well-known asthenoteratozoospermia-related gene, presents quite similar defects with DNAH12, indicating DNAH12 and DNAH1 possibly exhibit synergic functions in axonemal proper arrangement (58–60). In addition to DNAH12 and DNAH1, deficiency in DNAH2 or DNAH6 also leads to high proportions of CP loss and disorganized axonemal structures (5, 54, 61). However, DNALI1, DNAH7, or DNAH10 deficiency results in no obvious CP loss but impaired IDA integrity (7, 8, 22, 47). These studies indicate that specific IDA components, like DNAH12 and DNAH1, are required for the maintenance of CP integrity. Moreover, discrepancies observed in the phenotypes resulting from deficiencies in specific IDAs suggest that multiple dynein proteins are produced to support the intricate organization of sperm flagella. Additionally, these proteins may exhibit distinct characteristics and perform unique functions to some extent. Noteworthy, deficiency in RS head proteins RSPH1, RSPH9, or DNAJB13 were related to axonemal CP loss phenotype (48, 50, 62), indicating that DNAH12, and other IDAs, possibly regulate the stability of CP through the cooperation with RS head proteins.

The localization of DNAH12 presents dynamic changes during spermiogenesis. DNAH12 is initially expressed in the cytoplasm of the round spermatid, while it’s enriched in the manchette structure of the elongating spermatid and the tail of the elongated spermatid. The manchette and the sperm tail axoneme are two microtubular platforms that transport proteins to the developing head and tail (63). Manchette is a transient skirt-like structure and functions as a scaffold for morphological remodeling, contributing to the formation of the sperm head, neck, and tail (64). Here, we found that DNAH12 deficiency results in abnormal manchette structure and compromised sperm head and tail morphology. Additionally, in sperm of Dnah12^−/− males, several CP complex proteins like SPAG6, and SPEF2 are not properly assembled and the axoneme structure of the tail was severely disorganized, indicating dynein components are responsible for proper axonemal components organization. Noteworthy, DNALI1, a potential DNAH12 interacting partner showed similar localization patterns to DNAH12, a study presented that DNALI1 recruits and stabilizes manchette component PACRG, and DNALI1 deficiency causes impaired sperm flagellar assembly (57). These data hint that dynein proteins like DNAH12 and DNALI, possibly work together in the manchette organization as well as sperm flagellar components assembly during sperm development.

DNAH12 was also detected in the lungs, tracheas, and female oviducts while no cilia morphology defects in these tissues or PCD symptoms were observed in both two mice models. Besides, none of the patients carrying bi-allelic DNAH12 mutations claimed symptoms of PCD via visits except for the reproductive problems. Dnah12^-/- female mice showed normal development and fertility and the mother (III:2) of patients from family 2, carrying homozygous DNAH12 mutation M2 had 6 offspring. Intriguingly, we found that the localization and abundance of DNALI1 and DNAH1 in oviductal or tracheal cilia were not affected after Dnah12 deletion, indicating that DNAH12 works differently in flagella and cilia. Deficiency of specific DNAH members like DNAH6 (65), or DNAH7 (22), but not DNAH2 (54), results in PCD symptoms. Hence, DNAH family members possibly present disparate processing of axonemal biogenesis and organization through evolution in flagella and cilia, further explorations could contribute to a more comprehensive understanding of the differences between flagella and cilia, and personalized clinical diagnosis of related diseases like male/female infertility and/or PCD (66).

Previously, two genetic screening studies mentioned unverified compound heterozygous mutations and a homozygous missense mutation in DNAH12, predicting their association with patients’ infertility. However, limited genetic evidence was provided, and the pathogenesis remained unclear (26, 27). Through investigations involving patients with DNAH12 variants and mutant mouse models, we provide direct genetic and molecular evidence that variants in DNAH12 are genetic causes of asthenoteratozoospermia, and DNAH12 along with DNAH1, DNALI1 is indispensable for sperm flagellar biogenesis. These data strongly support DNAH12 as a valuable marker in genetic counseling and diagnosis of male infertility.

ICSI is a widely used assisted reproductive technology in treating infertile men with asthenoteratozoospermia. Deficiency in most dynein proteins leads to asthenoteratozoospermia with severely compromised sperm motility and morphology. Previous studies have shown that patients carrying mutations in DNAH17 (10) have poor ICSI outcomes while patients with DNAH1(59), DNAH2 (5, 61), DNAH3 (67), DNAH7 (22), or DNAH8 (68) mutations have achieved favorable clinical pregnancy. Thus, the comparative study of asthenoteratozoospermia-related genes on their ICSI outcomes will be beneficial to developing individualized treatment recommendations in clinical practice. In our study, the successful ICSI outcomes of Dnah12^−/− mice suggest that ICSI could be used as a potential treatment for infertile men carrying Dnah12 mutations. It’s meaningful to evaluate the prognosis for ICSI treatments in patients carrying DNAH12 variants in a large population.

In conclusion, our study functionally and mechanistically sheds light on a special and poorly characterized protein DNAH12, during the intricate process of sperm flagellar biogenesis. These results pave the way for an enhanced understanding of the roles of dynein proteins in sperm flagellar development and their functional difference in flagella and cilia, offering a diagnostic target for genetic counseling and future individualized treatment of male infertility in clinical practice.

Data availability

All data that support the findings in this study are available from the corresponding author upon reasonable request. Proteomics data supporting this study have been deposited to the ProteomeXchange Consortium via the iProX partner repository with the dataset identifier PXD051681.

Author contributions

B.S. and Q.S. conceived and supervised the study; M.Y., X.H., J.Y, M.C., A.Z., T.L., A.Z., and N.A. performed the experiments and analyzed the data; H.M.J, M.K., Z.M., R.K., A.A., and W.S. collected the patients’ samples; J.Z. and H.Z. performed the WES sequencing and data analyses; M.Y. wrote the manuscript and B.S., Q.S., H.M., and M.Y. modified the manuscript; B.S., Q.S, H.M., B.X., H.Z., and A.M. were involved in discussions of the manuscript.

Ethics statement

This work was approved by the institutional ethics committee of University of Science and Technology of China (USTC) with the approval number 2019-KY-168, and informed consent was obtained from patients and controls. All animal experiments were approved by the Institutional Animal Care and Use Committee of the USTC with the approval number USTCACUC1301021.

Funding

This work was supported by the National Key Research and Development Program of China (2022YFC2702601, 2022YFA0806303, and 2019YFA0802600), the National Natural Science Foundation of China (82171599, 32270901, and 32070850), the Global Select Project (DJK-LX-2022010) of the Institute of Health and Medicine, Hefei Comprehensive National Science Center, the Joint Fund for New Medicine of USTC (YD9100002034) and Scientific Research Foundation for Scholars of the First Affiliated Hospital of USTC (RC2023054).

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

We are grateful to all the participants for their cooperation. We appreciated Li Wang and Dandan Song at the Center of Cryo-Electron Microscopy, Zhejiang University, for their technical assistance with TEM and SEM, respectively. We thank the Bioinformatics Center of the USTC, School of Life Sciences, for providing supercomputing resources.